Yacc++ and the Language Objects Library

"We have had Yacc++ for a number of years now. We have been developing a distributed development architecture based on Microsoft COM / DCOM. From Monday 16th November, 1998 this product will be available for Delphi and soon after will be available for C++ Builder and Visual Basic. We also have plans for Visual C++ and Visual J++ releases in the first quarter of next year. The distributed development architecture consists of over 30 integrated components for the development of distributed applications. These components support identification and publication of distributed objects, navigation and resolution of objects, event notification and subscription, data streaming, expressions, security, testing and tracing.

One of the components in the distributed development architecture is a Language Manager component which provides distributed expressions, sets, identifiers and commands. The Language Manager creates COM objects by using a Yacc++ BNF and a novel object resolution mechanism that we have devised.

We have found the quality of the Yacc++ product to be excellent. The two best features from my point of view have been the tutorials which allowed me to make the transition from yacc / lex (and bison) to Yacc++ and the extensibility of its implementation. A constant source of frustration with other software development tools has been the lack of "openness" i.e. the ability to extend the capabilities of the product. This has definitely not been the case with Yacc++ - we have used numerous features supplied with Yacc++ (such as the ability to switch lexical analysers midstream) and have added other capabilities necessary for our product.

Early next year we plan to release our Knowledge Manager tool. This product allows knowledge to be represented, maintained and accessed in the form of decision tables. It also provides an editor for viewing, manipulating and testing knowledge. The Knowledge Manager integrates heavily with the Language Manager and the distributed development infrastructure to provide applications with the ability rapidly change knowledge without application recompilation."

"We have been using Yacc++ in two projects since 1993, and think it deserves great praise. The product is nicely designed, robust and fast; the documentation is well written, to the point, and complete; and the support is immediate and very competent."

"After using lex/yacc on Unix and then moving to Yacc++ on Windows NT, there's no comparison at all. Yacc++ provides many additional support methods and parsing/lexing shortcuts that make it a joy to use compared to lex/yacc. When you throw in the fact that it is a full C++ class library, you've got a very powerful tool."

"The Venus project, a research program from the Department of Computer Sciences at the University of Texas at Austin, is aimed at dramatically improving the execution speed of forward chaining rule based programs (e.g. expert systems). Whereas all existing production system environments are interpreted, a fundamental assumption of Venus is that rule based programs can be compiled. This allows significant amounts of work to be moved from the execution time to compile time.

The Venus project has been steadily evolving since its origination in 1992. At that time, the core of Venus was created in the form of a compiler. The compiler was created using proper design principles, and featured distinct front end and back end modules, with a graph based intermediate representation.

The initial version supported a rule language based on LISP. The compiler front end was constructed using YACC++. Later, a requirement emerged to support a rule language based on C++ syntax. A new compiler front end was written in YACC++ that translates C++ style rules into the previously defined intermediate representation. Thus, for several years, we have been able to maintain compilers for both LISP style and C++ style input languages with a minimum of incremental effort."